The UK Prospective Diabetes Study (UKPDS) is a prospective randomized study of different therapies that was set up to determine whether improved blood glucose control of non-insulin dependent diabetes will reduce morbidity and mortality. 5,102 patients have been recruited with median age 52 years, median fasting plasma glucose 12 mmol/l. In order to gain sufficient power, the clinical study will continue to 1998 with 12 years median follow-up. The allocation to 'active therapy' with either sulphonylurea or insulin has provided improved blood glucose control compared with allocation to 'diet policy' (median fasting plasma glucose 7.0 and 8.9 mmol/l and haemoglobin Alc 6.5 and 7.2% respectively). Therapy with metformin is also being studied. Since 1982, patients have had their fundi assessed by quantitation of retinal photographs with a modified '191' grading system at diagnosis and subsequently at three yearly intervals. At diagnosis of diabetes. 30% of patients have retinopathy. defined as microaneurysms (20 10 or worse), and 8% more severe retinopathy (31 31 or worse). Retinopathy defined as grade 31 31 or worse increased in prevalence to 13% at 3 years and to 17% at 6 years. Already 244 patients have had photocoagulation. 147 patients have had cataract extraction and 90 patients have been registered blind. The main analysis will be to determine whether improved blood glucose control will reduce the progress of retinopathy, assessed by deterioration of the '191' grading, and the occurrence of photocoagulation, cataract extraction and blindness. The study will examine both primary prevention and secondary prevention of diabetic retinopathy. In addition, several potential risk factors are being quantitated, e.g. smoking, blood pressure, plasma lipids, microalbuminuria to assess whether these are associated with the progress of retinopathy. In 1987, in a factorial design, a prospective randomized trial of different blood pressure therapies was included to determine whether improved blood pressure control will prevent complications of diabetes. 1,148 hypertensive patients have been allocated to tight blood pressure control with either a beta-blocker or ACE inhibitor, or to less-tight control. A satisfactory blood pressure difference has been obtained between the two groups (143/82 and 157/89 respectively). The study will also determine whether improved blood pressure control will prevent progress of retinopathy. Since ACE inhibitors are reported to prevent progress of diabetic renal disease, the retinal response to ACE inhibitors is of specific interest. The application is for the final retinal photographs, including those at the last triennial review in each patient, to determine over 6-19 years whether any specific therapeutic approach is advantageous and whether there is a threshold for glycaemia below which patients are protected. In Type 2 diabetic patients the effect of different degrees of glycemic control below those obtained in therapeutic studies of Type 1 patients can be studied, and may provide data in the near-normoglycemia range that will not otherwise be available.